The Art of Responsible Innovation in Financial Services

Miranda Jones, SVP | Data & AI Strategy Leader, Emprise Bank

Miranda Jones is SVP and Data & AI Strategy Leader at Emprise Bank. Blending a foundation in mathematics with a passion for people, she leads with integrity; building trust, empowering teams and shaping responsible, human-centered innovation in financial services.

Early Lessons that Shaped a Human-Centered Leadership Approach

Many of the projects that my team has supported have really challenged the status quo in how other teams at our bank use data and in turn how they work and make decisions. There are some key experiences early in my career that have shaped how I approach leading teams through this change.

When I was in graduate school doing research in theoretical mathematics, my thesis advisor would challenge me, saying “If you cannot draw a picture of this, you do not understand it.” Then, as a graduate teaching assistant and high school teacher, I was constantly reframing and breaking down math problems that my students were learning to solve. In many cases, I taught them to draw pictures to help them grasp some of the more abstract concepts.

Also, as an entry level analyst, I saw teams that had little advocacy or input on changes they were experiencing really fighting against the change. Outcomes could have been so different if their managers leaned in and made them part of the change. These experiences as a learner, educator and change leader have shaped my focus to being very intentional on bringing teams along, leveraging their knowledge and experience, helping them to give input and direction to the projects, and painting a picture of what the future could be.

Laying the Groundwork for Data-Driven Culture

Early on in my time at Emprise, I spent time learning from internal stakeholders about how they used data and what their expectations were for predictive analytics and dashboards. It was clear that not every team was at the same readiness level for machine learning. I focused on meeting the teams where they were to ensure that we could help them be more data-driven before trying to incorporate machine learning and AI into their work. Other teams were very eager to experiment and try using machine learning, so we focused early projects with their teams. Data-driven transformation is not a one size fits all process.

Balancing Automation with Human Oversight

One of the keys has been to focus on the business problem that we are solving. When solving these problems, we very closely partner with our internal stakeholders to define what model learning metric is most appropriate for the problem and what the expectations are for a minimum viable product. Beyond that, we lean on the subject matter experts to help us determine what the best variables are to test and we are transparent in our process of model training and monitoring. We want our models and systems to be trusted and used. Integrity is one of our core values; it is part of our DNA. We create transparent documentation and monitoring of model learning metrics and business metrics beyond what baseline requirements are for compliance.

“Integrity is one of our core values; it is part of our DNA. We create transparent documentation and monitoring of model learning metrics and business metrics beyond what baseline requirements are for compliance.”

Human-in-the-loop design, whether it is within an AI agent’s coded tasks or is set up to enable visibility of outcomes of a group of AI agents or fully automated machine learning pipelines, will be an on-going expectation. The amount of involvement of a person should be determined by the overall risk of the system and the impact it has when AI agents make mistakes.

Redefining Fast Innovation in a Regulated World for the Next Generation

We have designed our processes and standards with transparency, oversight and governance as core pillars. We have high expectations in our documentation to not only explain why we built the system and used the algorithm(s) we did, but also how to fix problems if any arise. This actually makes us faster when we are training up new team members or making decisions for future projects.

AI and analytics leaders need to focus on learning about their customers, internal or external. They need to understand who they are creating models, dashboards and AI systems for and for what purpose. I believe key differentiated skills include the ability to translate complex methods to business problems, ask good questions, reframe business problems into problems their teams can solve with clear goals, and be a curious learner. Being able to visualize, and maybe even draw a picture of, the AI system and the way it enables teams and organizations will also help bring others along. 

AI

Machine Learning

Financial

Predictive Analytics

Test